2 files changed, 62 insertions, 5 deletions
diff --git a/clang/test/CodeGen/char-literal.c b/clang/test/CodeGen/char-literal.c
index 322041c0049..014f6eb4fb0 100644
--- a/clang/test/CodeGen/char-literal.c
+++ b/clang/test/CodeGen/char-literal.c
@@ -1,5 +1,5 @@
-// RUN: %clang_cc1 -x c++ -triple i386-unknown-unknown -emit-llvm %s -o - | FileCheck %s
-// Runs in c++ mode so that wchar_t is available.
+// RUN: %clang_cc1 -x c++ -std=c++0x -triple i386-unknown-unknown -emit-llvm %s -o - | FileCheck %s
+// Runs in c++0x mode so that wchar_t, char16_t, and char32_t are available.
 
 int main() {
   // CHECK: store i8 97
@@ -16,6 +16,20 @@ int main() {
   // CHECK: store i32 98
   wchar_t wb = L'ab';
 
+  // CHECK: store i16 97
+  char16_t ua = u'a';
+
+  // Should pick second character.
+  // CHECK: store i16 98
+  char16_t ub = u'ab';
+
+  // CHECK: store i32 97
+  char32_t Ua = U'a';
+
+  // Should pick second character.
+  // CHECK: store i32 98
+  char32_t Ub = U'ab';
+
   // Should pick last character and store its lowest byte.
   // This does not match gcc, which takes the last character, converts it to
   // utf8, and then picks the second-lowest byte of that (they probably store
@@ -26,10 +40,36 @@ int main() {
   // CHECK: store i32 61451
   wchar_t wc = L'\uF00B';
 
+  // -4085 == 0xf00b
+  // CHECK: store i16 -4085
+  char16_t uc = u'\uF00B';
+
+  // CHECK: store i32 61451
+  char32_t Uc = U'\uF00B';
+
   // CHECK: store i32 1110027
   wchar_t wd = L'\U0010F00B';
 
+  // Should take lower word of the 4byte UNC sequence. This does not match
+  // gcc. I don't understand what gcc does (it looks like it converts to utf16,
+  // then takes the second (!) utf16 word, swaps the lower two nibbles, and
+  // stores that?).
+  // CHECK: store i16 -4085
+  char16_t ud = u'\U0010F00B';  // has utf16 encoding dbc8 dcb0
+
+  // CHECK: store i32 1110027
+  char32_t Ud = U'\U0010F00B';
+
   // Should pick second character.
   // CHECK: store i32 1110027
   wchar_t we = L'\u1234\U0010F00B';
+
+  // Should pick second character.
+  // CHECK: store i16 -4085
+  char16_t ue = u'\u1234\U0010F00B';
+
+  // Should pick second character.
+  // CHECK: store i32 1110027
+  char32_t Ue = U'\u1234\U0010F00B';
+
 }
diff --git a/clang/test/CodeGen/string-literal.c b/clang/test/CodeGen/string-literal.c
index cc6c0943d95..6d14330a0b9 100644
--- a/clang/test/CodeGen/string-literal.c
+++ b/clang/test/CodeGen/string-literal.c
@@ -1,4 +1,5 @@
-// RUN: %clang_cc1 -triple i386-unknown-unknown -emit-llvm %s -o - | FileCheck %s
+// RUN: %clang_cc1 -x c++ -std=c++0x -triple i386-unknown-unknown -emit-llvm %s -o - | FileCheck %s
+// Runs in c++0x mode so that wchar_t, char16_t, and char32_t are available.
 
 int main() {
   // CHECK: internal unnamed_addr constant [10 x i8] c"abc\00\00\00\00\00\00\00", align 1
@@ -9,8 +10,24 @@ int main() {
   char b[10] = "\u1120\u0220\U00102030";
 
   // CHECK: private unnamed_addr constant [12 x i8] c"A\00\00\00B\00\00\00\00\00\00\00", align 1
-  void *foo = L"AB";
+  const wchar_t *foo = L"AB";
 
   // CHECK: private unnamed_addr constant [12 x i8] c"4\12\00\00\0B\F0\10\00\00\00\00\00", align 1
-  void *bar = L"\u1234\U0010F00B";
+  const wchar_t *bar = L"\u1234\U0010F00B";
+
+  // CHECK: private unnamed_addr constant [12 x i8] c"C\00\00\00D\00\00\00\00\00\00\00", align 1
+  const char32_t *c = U"CD";
+
+  // CHECK: private unnamed_addr constant [12 x i8] c"5\12\00\00\0C\F0\10\00\00\00\00\00", align 1
+  const char32_t *d = U"\u1235\U0010F00C";
+
+  // CHECK: private unnamed_addr constant [6 x i8] c"E\00F\00\00\00", align 1
+  const char16_t *e = u"EF";
+
+  // This should convert to utf16.
+  // CHECK: private unnamed_addr constant [10 x i8] c" \11 \02\C8\DB0\DC\00\00", align 1
+  const char16_t *f = u"\u1120\u0220\U00102030";
+
+  // CHECK: private unnamed_addr constant [4 x i8] c"def\00", align 1
+  const char *g = u8"def";
 }